Response of Human Pulmonary Epithelial Cells to Lipopolysaccharide Involves Toll-like Receptor 4 (TLR4)-dependent Signaling Pathways

Pulmonary epithelial cells are continuously exposed to microbial challenges as a result of breathing. It is recognized that immune myeloid cells express Toll-like receptors (TLRs), which play a major role in detecting microbes and initiating innate immune responses. In contrast, little is known concerning the expression of TLR in pulmonary epithelial cells per se, their distribution within the cell, their function, and the signaling pathways involved. In this work, we demonstrated by reverse transcription-PCR and/or immunoblot that TLR4 and the accessory molecule MD-2 are constitutively expressed in distinct human alveolar and bronchial epithelial cells. We further characterized by flow cytometry, biotinylation/precipitation, and confocal microscopy the intracellular localization of TLR4 in these cells. Despite this intracellular compartmentalization of TLR4, pulmonary epithelial cells were responsive to the TLR4 activator lipopolysaccharide (LPS), a potent Gram-negative bacteria-associated molecular pattern. Using respiratory epithelial cells isolated from TLR4 knock-out and wild type mice, we demonstrated that TLR4 is the actual activating receptor for LPS in these cells. Furthermore we showed that this cell response to LPS involves a signaling complex including the kinases interleukin-1 receptor-associated kinase (IRAK), p38, Jnk, and ERK1/2. Moreover, using vectors expressing dominant-negative forms of MyD88 and TRAF6, we established that LPS-induced activation of respiratory epithelial cells is largely dependent on TLR4 signaling intermediates. Altogether these data demonstrate that TLR4 is a key element in the response of pulmonary epithelial cells to molecules derived from Gram-negative bacteria. The intracellular localization of TLR4 in lung epithelia is expected to play an important role in the prevention of the development of chronic inflammatory disease.

Pulmonary epithelial cells are continuously exposed to microbial challenges as a result of breathing. It is recognized that immune myeloid cells express Toll-like receptors (TLRs), which play a major role in detecting microbes and initiating innate immune responses. In contrast, little is known concerning the expression of TLR in pulmonary epithelial cells per se, their distribution within the cell, their function, and the signaling pathways involved. In this work, we demonstrated by reverse transcription-PCR and/or immunoblot that TLR4 and the accessory molecule MD-2 are constitutively expressed in distinct human alveolar and bronchial epithelial cells. We further characterized by flow cytometry, biotinylation/precipitation, and confocal microscopy the intracellular localization of TLR4 in these cells. Despite this intracellular compartmentalization of TLR4, pulmonary epithelial cells were responsive to the TLR4 activator lipopolysaccharide (LPS), a potent Gram-negative bacteria-associated molecular pattern. Using respiratory epithelial cells isolated from TLR4 knock-out and wild type mice, we demonstrated that TLR4 is the actual activating receptor for LPS in these cells. Furthermore we showed that this cell response to LPS involves a signaling complex including the kinases interleukin-1 receptor-associated kinase (IRAK), p38, Jnk, and ERK1/2. Moreover, using vectors expressing dominant-negative forms of MyD88 and TRAF6, we established that LPS-induced activation of respiratory epithelial cells is largely dependent on TLR4 signaling intermediates. Altogether these data demonstrate that TLR4 is a key element in the response of pulmonary epithelial cells to molecules derived from Gram-negative bacteria. The intracellular localization of TLR4 in lung epithelia is expected to play an important role in the prevention of the development of chronic inflammatory disease.
The lung is constantly exposed to invading particles and potential pathogens. To cope with this pressure, the lung has evolved a sophisticated defense mechanism designed to clear offending agents while inducing a minimum amount of concomitant inflammation. At first, mechanical defenses constituted by the mucociliary escalator participate in the removal of material from the tracheobronchial tree (1). Then resident and recruited phagocytes in the lower respiratory tract and alveoli ingest particulate matter and pathogens that circumvent this first line of defense. Pulmonary epithelial cells also maintain mucosal integrity by modulating local immune responses. Thus, these cells respond to a range of stimuli by producing biologically active mediators, including cytokines and chemokines, that influence airway inflammation (2).
In professional immune cells, including monocytes, macrophages, T cells, and dendritic cells, many receptors participate in microbe detection. Toll-like receptors (TLRs) 1 represent a conserved family of innate immune recognition receptors that play key roles in detecting microbes, initiating innate immune responses, and linking innate and adaptive immunity (3,4). Among these stimuli, the cell wall of Gram-negative bacteria contains lipopolysaccharide (LPS), a potent proinflammatory pathogen-associated molecular pattern. The response to LPS is initiated upon its interaction with TLR4 in conjunction with the accessory molecules MD-2 and soluble or membrane-bound CD14 (5). The response is then transduced via the interleukin (IL)-1 receptor signaling complex, which includes two essential adaptor proteins, myeloid differentiation (MyD)88 and tumor necrosis factor receptor-associated factor (TRAF)6 as well as the serine-threonine kinase known as IRAK. Other components involved in this signaling pathway include mitogen-activated protein kinases (MAPKs) such as extracellular signal-regulated kinase 1/2 (ERK1/2), c-Jun N-terminal kinase (Jnk), and p38 kinase (p38) (6,7). This signal transduction pathway further coordinates the induction of multiple genes encoding inflammatory mediators and co-stimulatory molecules (8).
Various studies have provided evidence that TLR4 plays a critical role in myeloid cells (9 -11), but recent reports suggest that an LPS signaling system also exists in cells of epithelial origin. Thus, TLR4 is expressed in intestinal (12)(13)(14)(15), renal (16), colonic, and gingival epithelia (17). However, the signaling pathways involved in the activation of these epithelia by LPS remain poorly defined. Furthermore little is known concerning the expression of TLR4 (18,19), its distribution within the cell, its function, and the corresponding signaling pathways in pulmonary epithelial cells despite the fact that these cells are remarkable for their critical physiologic exposure to frequent airborne microbial challenges.
Here we demonstrate that TLR4 is constitutively expressed in distinct human alveolar and bronchial epithelial cells and describe the intracellular localization of this receptor. We also establish that LPS-induced stimulation of these cells is dependent on the activation of TLR4 signaling intermediates such as MyD88, IRAK, TRAF6, and MAPK.
Isolation and Primary Culture of Murine Pulmonary Epithelial Cells-Pulmonary epithelial cells from pathogen-free C57BL/6 mice were isolated according to a modified protocol of Corti et al. (21). In brief, lungs were perfused with 10 ml of RPMI 1640 medium (supplemented with 1% antibiotics and 2 mM L-glutamine) through the pulmonary artery until they were cleared of blood. Bronchoalveolar lavages were performed using 5 ϫ 1 ml of PBS, 1 mM EDTA to remove alveolar leukocytes. One ml of 0.1% protease type XIV (Sigma) was instilled into the lungs through the trachea. Lungs were then removed and placed in a sterile tube containing 2 ml of RPMI 1640 medium (supplemented with 1% antibiotics and 2 mM L-glutamine) and incubated at 4°C overnight for enzymatic digestion to occur. Lung tissues were further teased apart in RPMI 1640 medium, 10% FCS in a Petri dish. The cell isolate was passed through a 100-m filter, and pulmonary epithelial cells were purified as follows. Collected cells were counted and incu-bated for 15 min at 4°C at the appropriate ratios with magnetic activated cell sorting (MACS) CD45 and CD90 microbeads (Miltenyi Biotech, Bergisch Gladbach, Germany), two antigens highly expressed by leukocytes and fibroblasts, respectively. Cells were then washed and diluted in PBS, 0.5% FCS, and CD45 minus /CD90 minus cells were negatively selected by passing the antibody-coated cell suspension through a column on an AutoMACS magnetic cell separator (Miltenyi Biotech). Pulmonary epithelial cells thus obtained were plated on Primaria 96well plate (4 ϫ 10 5 cells/well, BD Biosciences) in RPMI 1640 medium, 10% FCS. After 24 h, cells were stimulated for 6 h by 1 g/ml LPS, and KC and IL-6 levels were measured in the resulting supernatants.
RT-PCR-Cells were grown on a cell culture flask (Techno Plastic Products, Trasadingen, Switzerland), and total RNA was extracted using an RNeasy kit (Qiagen, Courtaboeuf, France). DNase treatment was performed using 4 g of extracted RNA, 1 l of DNase I (Amersham Biosciences), and 0.5 l of RNasin (Promega, Madison, WI) in a total volume of 20 l in the manufacturer's buffer. cDNA was obtained by incubating RNA with 1 mM dNTP (Eurobio, Les Ulis, France), 1.5 l of hexamers as primers, 20 units of RNasin (Promega), and 300 units of Moloney murine leukemia virus reverse transcriptase RNase H minus (Promega) in a total volume of 50 l in the manufacturer's buffer for 1 h at 42°C and 10 min at 70°C. PCR was performed using specific primers (Genset, Evry, France) for human TLR4 (sense, 5Ј-TGG ATA CGT TTC CTT ATA AG-3Ј; antisense, 5Ј-GAA ATG GAG GCA CCC CTT C-3Ј) and for human MD-2 (sense, 5Ј-GTT ACT GAT CCT CTT TGC ATT TGT AAA GCT TTG GAG-3Ј; antisense, 5Ј-TCT AGA CTA ATT TGA ATT AGG TTG GTG TAG GAT GAC AAA C-3Ј) (14,22,23). As an internal control, we used primers for the detection of human ␤-actin (sense, 5Ј-AAG GAG AAG CTG TGC TAC GTC GC-3Ј; antisense, 5Ј-AGA CAG CAC TGT GTT GGC GTA CA-3Ј). Amplifications were performed in a Peltier thermal cycler apparatus (MJ Research, Watertown, MA) using Q-BioTaq TM polymerase (Qbiogene, Illkirch, France). For the detection of TLR4 and MD-2, the thermocycling protocol was as follows: 95°C for 3 min, 36 cycles of denaturation at 95°C for 45 s, annealing at 56°C for 45 s, and extension at 72°C for 1 min. For the detection of ␤-actin, only 30 cycles were applied, and the temperature of annealing was 62°C. Amplification products were resolved on a 1.5% agarose gel containing ethidium bromide. Gels were recorded after amplification with an Ultra-Lum system (Ultra-Lum, Claremont, CA).
Immunoblotting-Epithelial and monocytic cells were washed once with cold PBS and then lysed on ice in a lysis buffer (10 mM Tris, 150 mM NaCl, 3 mM EDTA) supplemented with protein inhibitors (100 M leupeptin, 10 M aprotinin, 20 g/ml soybean trypsin inhibitor, 1 mM phenylmethylsulfonyl fluoride, 5 mM benzamidine) and 1% (v/v) Triton X-100. Samples were further solubilized prior to electrophoresis by adding SDS (2%, v/v), and disulfide bonds were reduced with 5% (v/v) ␤-mercaptoethanol. An equal amount of protein (15 g) was fractionated by SDS-PAGE on a 10% acrylamide gel, and proteins were further electrotransferred to a nitrocellulose membrane (Optitran BA-S 85, Schleicher & Schuell) and probed by immunoblotting using specific antibodies as specified in the figure legends. Bound antibodies were detected using the ECL ϩ immunoblotting detection system (Amersham Biosciences) according to the manufacturer's instructions. Concerning the biotinylation experiments, TLR4 expression at the cell surface was assessed by labeling A549, BEAS-2B, and U-937 cells with 1 mg/ml cell-impermeant sulfo-NHS-LC-biotin for 30 min. Then cells were incubated with PBS, 50 mM glycine to stop the reaction, washed three times with PBS, and solubilized with the above lysis buffer for 30 min at 4°C. After centrifugation for 30 min at 18,000 ϫ g at 4°C, the total protein concentration was determined using the Pierce protein assay, and 100 g of protein were immunoprecipitated for 1 h at 4°C with neutravidinagarose beads. After centrifugation, supernatants, corresponding to cytosolic fractions, were collected. Next beads were washed four times with the lysis buffer, and proteins were eluted using 2% SDS and further resolved by SDS-PAGE on a 10% acrylamide gel. After electrotransfer to a nitrocellulose membrane, TLR4 was detected using a specific antibody (H80, 0.5 g/ml). CD87 expression was also assessed under the same conditions to validate the biotinylation protocol as this receptor is known to be expressed both in the intracellular and surface compartments of pulmonary epithelial cells (24). Molecular masses were considered with respect to calibration standards included in each gel.
IRAK Immunoprecipitation and Kinase Assay-BEAS-2B cells were stimulated by 1 g/ml LPS for 0, 5, 15, 30, and 60 min. Cells were washed with cold PBS and lysed on ice in lysis buffer (50 mM HEPES, pH 7.6, 150 mM NaCl, 1 mM EDTA, 20 mM ␤-glycerophosphate, 0.1% Nonidet P-40, 1 mM sodium orthovanadate, 1 mM sodium fluoride, 1 mM dithiothreitol, and protease inhibitors). Cell debris were pelleted by centrifugation at 18,000 ϫ g at 4°C for 15 min. The protein concentration in the supernatants was determined according to the Pierce method. Equal amounts of extracts were incubated with 1 g of a polyclonal rabbit anti-IRAK antibody (Bio Vision Research Products, Palo Alto, CA) for 2 h at 4°C on a rotor. Fifty microliters of 50% protein G-Sepharose (Amersham Biosciences) were then added to each sample and incubated for an additional 2 h at 4°C. Samples were centrifuged briefly in a microcentrifuge, and the beads were washed once with lysis buffer and twice with kinase buffer (20 mM HEPES, pH 7.6, 20 mM MgCl 2 , 20 mM ␤-glycerophosphate, 20 mM p-nitrophenyl phosphate, 1 mM EDTA, and 1 mM benzamidine). The beads were incubated for 30 min at 30°C in a final volume of 20 l in the presence of 2 g of myelin basic protein (Sigma), 100 M ATP, and 5 Ci of [␥-32 P]ATP (PerkinElmer Life Sciences). SDS sample buffer was then added to protein G beads and boiled for 5 min. The samples were subjected to SDS-PAGE analysis. The gels were dried, and the intensity of the radioactive signal was quantified using a PhosphorImager (Amersham Biosciences) and NIH Image Version 1.62 software.
Fluorescence-activated Cell Sorter Analysis-Epithelial and monocytic cells were dispensed at 1 ϫ 10 6 cells/ml in conical bottom 96-well plastic plates (Nunc A/S, Roskilde, Denmark) and were further centrifuged at 100ϫ g at 4°C for 10 min. Cells were washed with Hanks' balanced salt solution, 0.5% bovine serum albumin supplemented with 1 mM Ca 2ϩ and Mg 2ϩ before adding a saturating concentration of HTA125 antibody (5 g/ml) or with non-immune IgG as control isotypes for 30 min at 4°C. Cells were further washed and incubated for 30 min at 4°C with the corresponding secondary FITC-conjugated antibody (10 g/ml). For intracellular staining, cells were fixed and permeabilized by incubating cells for 90 min on ice with a solution of PBS, 3.2% paraformaldehyde, 0.2% Tween 20 prior to incubation with a primary anti-TLR4 antibody (H80, 5 g/ml) and FITC-conjugated secondary antibodies (10 g/ml). Finally plates were centrifuged as above, cells were resuspended in the same buffer, and fluorescence analysis was performed using a FACScan flow cytometer (BD Biosciences).
Immunostaining and Confocal Microscopy Analysis-Both human primary bronchial epithelial cells and alveolar and bronchial epithelial cell lines were used for these experiments. A549 and BEAS-2B epithelial cells were cultured on 22-mm glass cell culture coverslips (CML France, Nemours, France). Cells were washed three times with a solution of PBS and then fixed for 15 min in a PBS, 3.2% paraformaldehyde solution. After washing under gentle shaking, cells were permeabilized for 5 min with 0.1% Triton X-100 and washed, as before, prior to incubation with primary and secondary antibodies as described in the figure legends. In control experiments, cells were incubated with nonimmune IgG as control isotypes. Rhodamine-phalloidin staining was used to visualize the actin cytoskeleton. Finally the cells were washed extensively with PBS, and the coverslips were mounted in fluorescence mounting medium. Primary epithelial cells were obtained by brushings performed on lobar bronchi of surgical lung resections. The lobectomy was performed in a context of peripheral subpleural malignant tumor. The proximal bronchial mucosa was evaluated as normal on gross examination. Each brushing was accomplished by very light gliding along the surface of the airway with a 1-mm cytology brush. The brushing sample was immediately placed in RPMI 1640 medium. Cells were then gently set down on coverslips by gentle cytocentrifugation and further processed as described above. Confocal microscopy was performed with a 63ϫ/1.4 oil objective lens on a confocal microscope (model LSM 510, Carl Zeiss France, Le Pecq, France) using laser excitation at 488 and 543 nm. The empty plasmid pcDNA3 (Invitrogen) was used to maintain the total plasmid quantity constant at 600 ng. After 24 h, cells were left untreated or stimulated for 6 h at 37°C with 1 g/ml LPS or 50 ng/ml TNF␣. Luciferase activity was measured in the cell lysates as described previously (27) using an EGNG Berthold luminometer. Results are expressed as relative luciferase units.
Statistical Analysis-Statistical significance between the individual groups was analyzed using the unpaired Student's t test with a threshold of p Ͻ 0.05.

Identification of Components of the LPS Receptor Complex in
Pulmonary Epithelial Cells-TLR4 is the primary signal transducer involved in LPS signaling (5). RT-PCR was used to assess the expression of this receptor in distinct unstimulated human respiratory epithelial cells. Phorbol 12-myristate 13-acetatedifferentiated U-937 cells (macrophage-like cells) have been shown previously to express TLR4 and were used as a positive control for this receptor (Fig. 1, lanes 1). As shown in Fig. 1, left panel, TLR4 is clearly expressed at the mRNA level in both human alveolar (A549) and tracheobronchial (BEAS-2B, 16-HBE, and NT-1) epithelial cell lines. A high level of expression is also observed in CFT-2, a human tracheal cell line with several phenotypic characteristics of cystic fibrosis respiratory epithelia (20). RT-PCR analysis of ␤-actin expression confirmed the quality of all mRNA preparations used for RT-PCR. Human TLR4 is also detectable in all five epithelial cell lines as an Ϸ110-kDa protein (Fig. 1, middle panel). An immunoblot for ␤-actin demonstrated similar gel loading. The apparent discrepancy between mRNA and protein expression of TLR4 is consistent with previous studies using intestinal epithelial cells (28) and suggests that expression of this receptor at the mRNA level may not be directly correlated with its protein level.
Current research findings suggest that the activation of the innate immune system involves complex associations of receptors depending on cell types and bacterial stimuli. The LPS receptor complex of mononuclear phagocytes is composed of TLR4, MD-2, and CD14. To gain insight into the mechanism of LPS recognition by pulmonary epithelial cells, we first ana- lyzed the expression of the accessory molecule MD-2 by the two cell lines A549 and BEAS-2B. Using RT-PCR, we observed that mRNA for MD-2 is expressed in these epithelial cells (Fig. 1,  right panel). The presence of the MD-2 protein was not checked due to the lack of a specific antibody. Concerning the presence of CD14, this molecule was detected by flow cytometry at a low expression level at the surface of BEAS-2B cells (not illustrated). By contrast, CD14 protein staining could not be detected either on A549 cells (not shown) or in primary polarized bronchial epithelial cells (see Fig. 4, lower panels). Intracellular Localization of TLR4 in Pulmonary Epithelial Cells-We first demonstrated by flow cytometry the absence of cell surface staining of TLR4 using A549 and BEAS-2B cells (Fig. 2, left panels). Positive staining of the monocytic cells U-937 confirmed the quality of the anti-TLR4 antibody and the flow cytometry protocol used (Fig. 2, inset). By contrast, using a mild fixation and permeabilization protocol, we identified significant levels of cytoplasmic TLR4 in pulmonary epithelial cells (see Fig. 2, right panels). As an independent measure of the intracellular presence of TLR4 in these cells, surface proteins were labeled by a membrane-impermeable form of biotin. Biotinylated proteins were then extracted with neutravidinagarose beads followed by an immunoblot with an anti-TLR4 antibody. Fig. 3A confirms that TLR4 is only detectable in the total or in the non-precipitated cytoplasmic fraction but not in the precipitated cell surface portion of respiratory epithelia. An immunoblot using horseradish peroxidase-conjugated streptavidin was also used to control the efficiency of the precipitation protocol (not shown). As an additional positive control of our biotinylation protocol, biotinylated proteins extracted from A549 and BEAS-2B cells were subjected to an immunoblot with an anti-CD87 antibody. Consistent with previous studies (24), we demonstrated that the CD87/urokinase plasminogen activator receptor is expressed both on the cell surface and the cytoplasmic compartments of pulmonary epithelial cells. Interestingly Fig. 3A also shows that TLR4 is clearly detectable both on the plasma membrane and intracellular fractions of U-937 cells. This observation is in good agreement with recent studies focusing on the distribution of TLR4 in human monocytes (29).
Furthermore immunolocalization by confocal microscopy confirmed the intracellular compartmentalization of TLR4. Staining in A549 and BEAS-2B cells exhibited a diffuse cytosolic pattern (Fig. 3B). This staining was not detected in samples in which the primary antibodies was omitted or in which a control irrelevant antibody was substituted for the primary antibody (Fig. 3B, right panels). More importantly, we demonstrated a similar intracellular, subapical pattern of TLR4 in human primary polarized bronchial epithelial cells (characterized by cilia on the apical cell surface visible using differential interference contrast microscopy (Fig. 4, upper panels)).
LPS Induces the Secretion of IL-8 and IL-6 but Not RANTES from Pulmonary Epithelial Cells-In view of the unexpected intracellular localization of TLR4 in respiratory epithelial cells, we considered whether these cells could nevertheless respond to LPS stimulation. Fig. 5, A and B, clearly shows that LPS purified from the respiratory pathogen P. aeruginosa strongly stimulated the release of IL-8 and IL-6 from BEAS-2B cells in a concentration-dependent manner in the presence of serum. This result is in agreement with previous studies (30). To rule out any stimulatory effect due to LPS contamination by other bacteria-derived components such as lipoproteins, experiments were also performed using LPS supplemented with 20 g/ml polymyxin B, a well characterized LPS inhibitor (31). Under these experimental conditions, IL-8 secretion by BEAS-2B cells were loaded on SDS-polyacrylamide gels. After transfer to nitrocellulose membranes, proteins were probed with an anti-TLR4 (H80, 0.5 g/ml) or an anti-CD87 antibody (clone 3932, diluted 1:10,000). The primary antibody was detected using a secondary antibody coupled to horseradish peroxidase (diluted 1:2000). B, shown are representative thin section confocal fluorescence micrographs of resting A549 and BEAS-2B cells labeled with anti-TLR4 or isotype control antibodies (5 g/ml). The primary antibody was detected using a secondary antibody coupled to Alexa488 (green, diluted 1:200). Rhodamine-phalloidin staining (red, diluted 1:2000) visualizes the actin cytoskeleton. Data are representative of three distinct experiments. Ab, antibody. was completely abolished (not shown). Interestingly LPS does not induce the release of RANTES, whereas this chemokine is clearly secreted upon stimulation of BEAS-2B cells by an optimal concentration of TNF␣ or IL-1␤ (Fig. 5C). To check whether TLR4 is the actual activating receptor for LPS, activation was assessed in respiratory epithelial cells from TLR4 knock-out mice in comparison with cells from wild type animals. Fig. 5D clearly shows that LPS strongly stimulates the release of KC and IL-6 in TLR4ϩ/ϩ pulmonary epithelial cells, while those mediators are absent in epithelial supernatants isolated from TLR4Ϫ/Ϫ animals.

LPS Signaling Does Not Modulate the Expression of TLR4 in Pulmonary Epithelial
Cells-Previous studies dealing with the regulation of TLR expression demonstrated contrasted findings depending on the stimulus, cell type, or tissue considered (32). We therefore examined whether LPS may regulate expression of TLR4 in respiratory epithelia by exposing BEAS-2B cells to an optimal concentration of LPS (1 g/ml, see Fig. 5) for 24 h. Expression of TLR4 mRNA and protein was normalized to ␤-actin and is reported as histogram bars. Based on this semiquantitative densitometric measurement, Fig. 6, A and B, clearly shows that LPS does not modulate TLR4 expression. Likewise incubation of BEAS-2B cells with LPS for a longer (up to 48 h) or shorter time (1-6 h) did not affect cellular levels of this receptor (not illustrated). Finally TLR4 is probably not recruited to the plasma membrane of pulmonary epithelial cells upon LPS activation. Indeed, when BEAS-2B cells were stimulated with LPS for various periods of time (5,15, and 60 min and 24 h), no TLR4 expression was found at the cell surface as assessed by flow cytometry analysis (not shown).
LPS Activates a TLR4-dependent Signaling Pathway in Pulmonary Epithelial Cells-To elucidate the signaling mecha-nisms associated with the activation of respiratory epithelia by LPS, we first examined the kinase activity of IRAK in BEAS-2B cells stimulated by 1 g/ml LPS. IRAK kinase activity was determined in vitro in cell extracts in the presence of [␥-32 P]ATP using myelin basic protein as a substrate. As shown in Fig. 7A, after a 15-min stimulation with LPS, myelin basic protein phosphorylation increased Ϸ2-fold as compared with unstimulated BEAS-2B cells. The kinase activity of IRAK in response to LPS was still detectable after a 1-h stimulation. We further assessed whether MyD88 and/or TRAF6 were involved in the NF-B signaling pathway activated by LPS using a reporter luciferase plasmid and either dominant-negative mutants or control vector. Transfection of BEAS-2B cells with increasing concentrations (50 -500 ng) of the expression vectors encoding MyD88-DN or TRAF6-DN resulted in a clear dose-dependent inhibition of LPS-mediated NF-B activation (Fig.  7B). As a control for inhibition specificity, TNF␣-induced NF-B activation, which is known to require TRAF2 (33), was not inhibited by 500 ng of either dominant-negative vector (not illustrated).
Many cellular stress stimuli are known to activate pathways associated with both NF-B and MAPKs such as p38, Jnk, and ERK1/2 (8,10). Because of this combination, we examined the activation of these MAPKs following LPS activation. This was determined by immunoblot analysis using anti-phosphorylated MAPK antibodies. Fig. 7C shows that in addition to NF-B, LPS strongly up-regulates p38, Jnk, and ERK1/2 activity in BEAS-2B cells. The highest level of MAPK phosphorylation occurred within 15-30 min and declined thereafter. DISCUSSION Due to inhalation of particles containing bacteria and LPS from the commensal flora in the nasopharynx and the environment, the lung is constantly exposed to potentially inflammatory components. Along with alveolar macrophages, pulmonary epithelial cells are the first cells to be challenged by LPS, and

FIG. 4. Evidence for an intracellular TLR4 localization in human primary bronchial epithelial cells.
Representative thin section micrographs showing localization of TLR4 in human primary bronchial epithelial cells by indirect immunofluorescence using an anti-TLR4 antibody (5 g/ml) and a secondary antibody coupled to Alexa488 (green, diluted 1:200). Isotype control antibodies were also tested to check the signal specificity. Cells were examined by differential interference contrast (left panels) and confocal microscopy (middle panels). The upper merged image (right panels) shows that TLR4 is localized in a subapical cytosolic compartment. Lower panels show no staining for the CD14 molecule using the specific FITC-conjugated anti-CD14 antibody (clone MY4, diluted 1:100). Ab, antibody; interfer., interference; fluo., fluorescence. their response must be greatly regulated to prevent alteration of the mucosal barrier. The recent discovery of TLR4 as the receptor for LPS prompted us to investigate the contribution of this receptor to the molecular mechanisms underlying the activation of pulmonary epithelial cells by this key product of Gram-negative bacteria. Two previous studies reported the expression of TLR4 mRNA in these cells (18,19). Here we confirm and extend these data by showing (i) that respiratory epithelial cells express a LPS receptor complex that includes TLR4 and MD-2, (ii) an unexpected intracellular compartmentalization of TLR4 that nevertheless allows LPS to strongly induce the secretion of proinflammatory mediators, (iii) that epithelial activation by LPS does not alter TLR4 expression at the mRNA or protein level or alter its intracellular localization, and (iv) that the LPS activation pathway shares elements with cells of the myeloid lineage in that TLR4-dependent signaling in epithelial cells involves intermediates such as MyD88, IRAK, TRAF6, and MAPK.
Secretion of inflammatory cytokines such as IL-8 and IL-6 by epithelial cells involves fairly elevated amounts (0.1-1 g/ml) of LPS in comparison with phagocytic cells, which are stimulated by 1-10 ng/ml LPS (10,11,22). Thus, initiation and coupling to downstream signaling events appear to be less efficient in the pulmonary epithelium than in myeloid cells. We hypothesize that this different responsiveness to LPS may be related to a distinct compartmentalization of TLR4 in the two cell types. Although apparently discordant data have been reported, an intracellular localization of TLR4 has been observed in epithelial cells of intestinal origin. Thus, while Cario et al. (13) reported cell surface expression of TLR4, others observed the absence of this receptor in the same type of epithelium (34). In the context of this discrepancy, a more recent study demonstrates the intracellular distribution of TLR4 in the Golgi apparatus where it co-localizes with internalized LPS and its absence on the surface of intestinal epithelial cells in contrast to its membrane expression on monocytes (15). Concerning pulmonary epithelial cells, by means of complementary techniques including flow cytometry, biotinylation/precipitation, and confocal microscopy using both cell lines and primary cells, we provide substantial evidence for an intracellular localization of TLR4 in these cells. Based on flow cytometry experiments, it is unlikely that LPS might recruit TLR4 to the cell surface upon cell activation. However, we cannot exclude the possibility that inflammatory mediators such as cytokines or bioactive lipids might be able to induce TLR4 relocalization. In agreement with the absence of TLR4 expression on the cell surface of pulmonary epithelial cells, it is also relevant to notice that addition of a blocking anti-TLR4 antibody (clone HTA125) in the extracellular medium had no effect on activation by LPS as assessed by the measurement of IL-8 secretion. By contrast, the same antibody reduced LPS-induced activation of U-937 cells by 68% (not shown). Thus, we speculate that the intracellular compartmentalization of TLR4 may prevent "inopportune" activation of pulmonary epithelial cells due to a regular FIG. 6. Expression of epithelial TLR4 is not regulated by LPS. BEAS-2B cells were stimulated or not (NS) for 24 h by 1 g/ml P. aeruginosa LPS. A, expression of human TLR4 and ␤-actin mRNA was analyzed by RT-PCR. B, TLR4 protein expression was assessed by immunoblot as described in Fig. 1, and membranes were further reprobed with an anti-␤-actin antibody (diluted 1:15,000) to confirm similar gel loading. Densitometric analysis was carried out using the NIH Image Version 1.62 software. Histogram bars show TLR4 normalized against ␤-actin expression and are representative of at least two distinct experiments.

FIG. 7. Activation of pulmonary epithelial cells by LPS involves TLR4-associated signaling pathways.
A, BEAS-2B cells were stimulated or not (NS) by 1 g/ml LPS, and the protein IRAK was further immunoprecipitated using a specific antibody as detailed under "Experimental Procedures." IRAK kinase activity was determined in vitro in the presence of [␥-32 P]ATP using myelin basic protein as a substrate. Samples were subjected to SDS-PAGE. The intensity of the radioactive signals was quantified using a PhosphorImager and NIH Image Version 1.62 software. Results are expressed as densitometric arbitrary values and are representative of two distinct experiments. B, BEAS-2B cells were transfected with increasing concentrations of expression plasmids for dominant-negative forms of MyD88 or TRAF6 and 100 ng of a NF-B-luciferase reporter construct. BEAS-2B cells were then stimulated by LPS (1 g/ml) for 6 h, and cell lysates were processed for luciferase activity. Results are expressed as relative luciferase units (RLU) and are the mean Ϯ S.D. of triplicate determinations of a representative experiment performed three times. C, time course of the activation of the MAPK p38, Erk1/2, and Jnk. BEAS-2B cells were stimulated or not (NS) by 1 g/ml LPS and further lysed, and phosphorylation of MAPK was determined by immunoblot using specific antibodies (all diluted 1:2000). To confirm equal loading, membranes were reprobed with an anti-␤-actin antibody (diluted 1:15,000). Data are representative of four distinct experiments. D, schematic model of the mechanisms involved in pulmonary epithelial cell activation by LPS. After a receptor-mediated active transfer or a passive delivery across the plasma membrane, LPS is internalized and rapidly transported to intracellular TLR4. Upon LPS-TLR4 interaction in a cytoplasmic compartment, downstream signaling is triggered, which involves at least the signal-transducing molecules MyD88, IRAK, TRAF6, MAPK, and the transcription factor NF-B. exposure to air containing trace amounts of LPS and as a consequence a chronic inflammatory state. In the context of this distinctive cell distribution, TLR4 signaling may therefore be triggered only upon exposure to a high amount of free or bacteria-associated LPS as occurs in occupational or infectious diseases (18,35). Subsequently the pulmonary epithelium may then participate in the local innate response through the secretion of cytokines and antimicrobial peptides.
Interestingly, before the identification of TLR4 as an essential participant in LPS signaling, Wright and colleagues (36,37) showed that LPS is rapidly delivered from the plasma membrane to an intracellular site and that agents that block vesicular transport alter cell responses to LPS. Moreover Vasselon et al. (38) demonstrated that monomeric LPS crosses the cell membrane and traffics within the cytoplasm independently of membrane CD14, while aggregates of LPS are internalized in association with CD14. In the present study, we attempted to immunolocalize CD14 in human primary polarized bronchial epithelial cells using confocal microscopy and a FITC-conjugated anti-CD14 antibody (clone MY4). In fact, no CD14 protein staining could be detected in lung epithelial samples. A similar result was observed using the pulmonary epithelial cell line A549 but was not seen with BEAS-2B cells, which express a low level of CD14. Thus, our data do not currently dissipate the debate that exists concerning the expression and role of CD14 in LPS-induced lung epithelial activation. Several authors proposed that these cells are CD14-negative (39,40), while others demonstrated both CD14 mRNA and cell surface protein in human airway epithelial cells (2,18,19). In fact, these contradictory results may be explained by distinct basal activation or differentiation state of the epithelial cells used throughout these investigations.
Regardless we may propose a hypothetical mechanism to construe how elevated concentrations of LPS activate human respiratory epithelial cells (see Fig. 7D). After a receptor-mediated active transfer or a passive delivery across the plasma membrane (36,37,41), LPS is internalized and rapidly encounters intracellular TLR4, and downstream signaling is triggered. Our present findings suggest that this process involves at least the signal-transducing molecules MyD88, IRAK, and TRAF6 and activation of the transcription factor NF-B. Also MAPKs appear to be important mediators of this cell activation process as three of these kinases (p38, Jnk, and ERK1/2) are selectively activated in a time-dependent manner by LPS.
In conclusion, our results suggest that resting respiratory epithelial cells constitutively express intracellular TLR4 and that secretion of inflammatory mediators upon exposure to LPS is likely a result of TLR4 signaling pathways. The peculiar compartmentalization of TLR4 is expected to play a critical role in the prevention of chronic pulmonary inflammatory disease. Of note is that it has become increasingly clear that a local dysfunction of innate immunity may result in lung inflammation as seen in cystic fibrosis. Thus, studies are currently underway to investigate whether TLR4 expression and cell distribution are altered in pulmonary mucosa of patients with cystic fibrosis or other lung disorders.